Alveolar bone fracture detection system, equipment, medium and program product

Abstract

The invention discloses an alveolar bone fracture detection system and device, a medium and a program product, and relates to the technical field of data identification, and the method comprises the steps: extracting multi-scale features of obtained oral image data, and segmenting to obtain a fracture region and a tooth root region; a central line is extracted from the fracture area, three-dimensional Euclidean distances are accumulated voxel by voxel along the central line to obtain the length of a fracture line, the central line is expressed as a function of the arc length, then the global average curvature is calculated, the number of branch points in the fracture area is detected, the minimum distance between the fracture area and the tooth root area is calculated, and therefore the fracture type is recognized according to the characteristic parameters; taking the mass center of the fracture area as a seed point, determining the gray value of the seed point, combining the newly added voxels which belong to the neighborhood expansion area of the original fracture area and have the difference value between the gray value and the gray value of the seed point meeting a set threshold value into the original fracture area to obtain an optimized fracture area, and calculating the area volume of the optimized fracture area. And the objectivity, accuracy and efficiency of diagnosis are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data identification, and in particular to an alveolar bone fracture detection system, equipment, medium and program product. Background Art

[0002] In recent years, intelligent algorithms, such as deep learning and image segmentation models, have made significant progress in medical imaging diagnosis, but the accurate diagnosis of alveolar bone fractures still faces the following technical bottlenecks.

[0003] (1) Insufficient adaptability to complex fracture morphologies.

[0004] (1-1) Existing models, such as convolutional neural networks and Mask R-CNN (Mask Region-based Convolutional Neural Network, a deep learning model for object detection and instance segmentation), are mostly designed based on regular shape targets, and have limited accuracy in segmenting the slender, multi-branched, and low-contrast fracture lines of alveolar fractures. For example, branch detection of comminuted fractures often results in breakage due to feature loss (false negative rate > 15%).

[0005] (1-2) In three-dimensional segmentation, traditional models (such as three-dimensional convolutional neural networks) rely on local context and have difficulty modeling the continuity of long-distance fracture lines, especially when the inter-layer resolution of CBCT (Cone beam CT) is insufficient, which easily produces "staircase artifacts".

[0006] (2) Challenges of small samples and domain generalization.

[0007] (2-1) Alveolar bone fracture imaging data are scarce and the labeling cost is high. Existing methods rely on large-scale labeled data training, which is prone to overfitting in small sample scenarios (such as rare periradicular fractures), and the generalization performance is significantly reduced (the accuracy of cross-center tests decreases by 20%-30%).

[0008] (2-2) Although transfer learning can alleviate the problem of insufficient data, the pre-trained model is very different from the medical imaging domain, and it is difficult to align the feature space, resulting in poor adaptability of the model to grayscale distribution and noise patterns.

[0009] (3) Defects in fusion of morphological and semantic features.

[0010] (3-1) Most studies rely only on end-to-end segmentation and lack explicit modeling of fracture morphology (length, curvature) and anatomical structure (root position), resulting in classification results that are out of line with clinical standards.

[0011] (3-2) The fusion strategy of multimodal features (such as grayscale, texture, and spatial coordinates) is single, and traditional concatenation or weighted methods are difficult to capture high-order interactions, which limits the classifier's discrimination ability.

[0012] (4) Error accumulation in three-dimensional volume measurement.

[0013] (4-1) Existing volume calculations are mostly based on the accumulation of layer-by-layer two-dimensional masks, ignoring the continuity between layers. The volume of inclined or spiral fractures can be underestimated by 10%-20%.

[0014] (4-2) Partial volume effect compensation methods (such as linear interpolation) fail at complex fracture boundaries, and region growing algorithms rely on fixed thresholds and are difficult to adapt to differences in bone density among different patients. Summary of the invention

[0015] In order to solve the above problems, the present invention proposes an alveolar bone fracture detection system, equipment, medium and program product, which can accurately locate the fracture line, automatically distinguish the fracture type, calculate the volume of the fracture area, and significantly improve the objectivity, accuracy and efficiency of diagnosis.

[0016] In order to achieve the above object, the present invention adopts the following technical solution: In a first aspect, the present invention provides an alveolar bone fracture detection system, comprising: A feature extraction module is configured to extract multi-scale features from the acquired oral image data, thereby segmenting the fracture area and the tooth root area; A type recognition module is configured to extract a center line of the fracture area, accumulate three-dimensional Euclidean distances voxel by voxel along the center line to obtain the length of the fracture line, express the center line as a function of the arc length and then calculate the global average curvature, detect the number of branch points in the fracture area, and calculate the minimum distance between the fracture area and the root area, thereby identifying the fracture type according to the length of the fracture line, the global average curvature, the number of branch points, and the minimum distance between the fracture area and the root area; The volume calculation module is configured to use the centroid of the fracture area as the seed point and determine the grayscale value of the seed point, merge the newly added voxels that belong to the neighborhood extension area of ​​the original fracture area and whose grayscale value difference with the grayscale value of the seed point meets the set threshold into the original fracture area, obtain the optimized fracture area, and calculate the regional volume of the optimized fracture area.

[0017] As an optional implementation, in the feature extraction module, in the process of extracting multi-scale features, an adaptive weight is calculated for each feature layer: ; in, It is The first step of oral imaging data Layer feature map; and They are The mean and standard deviation of the layer feature map; is the smoothing factor; It is Step Adaptive weights of layer features; for The spatial position of each voxel in; Multi-scale feature fusion based on adaptive weights: ; in, is the fusion feature, is the total number of feature layers.

[0018] As an optional implementation, in the type identification module, the fracture line length for: ; in, For the The coordinates of the centerline voxels; For the The coordinates of the centerline voxels; is the total prime number of the center line.

[0019] As an optional implementation, in the type recognition module, the process of calculating the global mean curvature includes: representing the center line as an arc length Function , the global mean curvature for: ; in, is the first-order derivative of the function; is the second-order derivative of the function; is the total prime number of the center line.

[0020] As an optional implementation, in the type recognition module, the process of detecting the number of branch points in the fracture area includes: traversing the skeleton points on the center line, determining whether there are connection directions greater than a set threshold number in the neighborhood of the current skeleton point; if so, the current skeleton point is regarded as a branch point, thereby determining the number of branch points in the fracture area.

[0021] As an optional implementation, in the type identification module, the process of identifying the fracture type includes: If the fracture line length is greater than or equal to the set length threshold, the global average curvature is less than or equal to the set first curvature threshold, and the number of branch points is less than or equal to the set first number threshold, then it is a linear fracture; If the number of branch points is greater than or equal to the set second number threshold, or the global average curvature is greater than the set second curvature threshold, it is a comminuted fracture; If the minimum distance between the fracture area and the root area is less than the distance threshold, it is a periradicular fracture.

[0022] As an optional implementation, in the volume calculation module, the centroid coordinates are ;in, are the coordinates of all fracture voxels in the fracture region binary mask, is the total number of all fracture voxels in the fracture area; Added Voxels The conditions for being merged are: ;in, is the gray value of the seed point, Tolerance, Neighborhood represents the neighborhood extension area of ​​the original fracture area; Area Volume for: ; in, The first The cross-sectional area of ​​the layer voxel; is the distance between oral image data layers; is the volume of the newly added voxels, and M is the number of newly added voxels.

[0023] In a second aspect, the present invention provides an electronic device, including a memory and a processor, and computer instructions stored in the memory and running on the processor, wherein when the computer instructions are run by the processor, a method for detecting alveolar bone fracture is completed, and the method for detecting alveolar bone fracture includes: Extract multi-scale features from the acquired oral image data, thereby segmenting the fracture area and the tooth root area; Extract the center line of the fracture area, accumulate the three-dimensional Euclidean distance voxel by voxel along the center line to obtain the length of the fracture line, express the center line as a function of the arc length, calculate the global average curvature, detect the number of branch points in the fracture area, and calculate the minimum distance between the fracture area and the root area. In this way, the fracture type can be identified based on the length of the fracture line, the global average curvature, the number of branch points, and the minimum distance between the fracture area and the root area. The centroid of the fracture area is taken as the seed point, and the grayscale value of the seed point is determined. The newly added voxels that belong to the neighborhood extension area of ​​the original fracture area and whose grayscale value difference with the grayscale value of the seed point meets the set threshold are merged into the original fracture area to obtain the optimized fracture area, and the regional volume of the optimized fracture area is calculated.

[0024] In a third aspect, the present invention provides a computer-readable storage medium for storing computer instructions, wherein when the computer instructions are executed by a processor, a method for detecting alveolar bone fracture is performed, wherein the method for detecting alveolar bone fracture comprises: Extract multi-scale features from the acquired oral image data, thereby segmenting the fracture area and the tooth root area; Extract the center line of the fracture area, accumulate the three-dimensional Euclidean distance voxel by voxel along the center line to obtain the length of the fracture line, express the center line as a function of the arc length, calculate the global average curvature, detect the number of branch points in the fracture area, and calculate the minimum distance between the fracture area and the root area. In this way, the fracture type can be identified based on the length of the fracture line, the global average curvature, the number of branch points, and the minimum distance between the fracture area and the root area. The centroid of the fracture area is taken as the seed point, and the grayscale value of the seed point is determined. The newly added voxels that belong to the neighborhood extension area of ​​the original fracture area and whose grayscale value difference with the grayscale value of the seed point meets the set threshold are merged into the original fracture area to obtain the optimized fracture area, and the regional volume of the optimized fracture area is calculated.

[0025] In a fourth aspect, the present invention provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, a method for detecting alveolar bone fracture is implemented, wherein the method for detecting alveolar bone fracture comprises: Extract multi-scale features from the acquired oral image data, thereby segmenting the fracture area and the tooth root area; Extract the center line of the fracture area, accumulate the three-dimensional Euclidean distance voxel by voxel along the center line to obtain the length of the fracture line, express the center line as a function of the arc length, calculate the global average curvature, detect the number of branch points in the fracture area, and calculate the minimum distance between the fracture area and the root area. In this way, the fracture type can be identified based on the length of the fracture line, the global average curvature, the number of branch points, and the minimum distance between the fracture area and the root area. The centroid of the fracture area is taken as the seed point, and the grayscale value of the seed point is determined. The newly added voxels that belong to the neighborhood extension area of ​​the original fracture area and whose grayscale value difference with the grayscale value of the seed point meets the set threshold are merged into the original fracture area to obtain the optimized fracture area, and the regional volume of the optimized fracture area is calculated.

[0026] Compared with the prior art, the present invention has the following beneficial effects: The present invention extracts multi-scale features from oral imaging data and segmentation obtains fracture areas and tooth root areas; in the process of extracting multi-scale features, an adaptive feature extraction mechanism is introduced. When processing each layer of features, the weights are automatically adjusted according to the significance of different areas. The model can be optimized according to the characteristics of different areas in the oral imaging data, making the detection of key areas (such as fracture lines) more sensitive, avoiding the excessive attention of traditional methods to unimportant areas. At the same time, by introducing a multi-scale feature fusion strategy, the low-level edge information and high-level semantic information are effectively combined, which not only improves the accuracy and robustness of image segmentation, but also enhances the model's ability to handle complex images (such as noise interference or edge fuzzy areas), and improves the diagnostic accuracy of fracture lines.

[0027] The present invention improves the multi-scale feature fusion mechanism of the segmentation model, introduces conditional random field optimization and dynamic region growing algorithm, accurately locates the fracture line, automatically distinguishes linear fractures, comminuted fractures, periradicular fractures and other types according to the fracture line morphology, position and relationship with the tooth root, accurately calculates the volume of the fracture area through three-dimensional reconstruction and region growing algorithm, integrates the diagnosis results, and outputs a quantitative report including the fracture type, position and volume, which significantly improves the objectivity, accuracy and efficiency of the diagnosis.

[0028] The present invention directly calculates the actual spatial volume based on the three-dimensional voxel model of the binary mask of the fracture area, avoiding the error caused by projection superposition in traditional two-dimensional measurement. It supports volume integration of complex fracture morphology (such as multi-region comminuted fracture), and the volume error compensation part can dynamically compensate for the missed volume of scattered fracture fragments. The regional growing algorithm is used for voxel refinement. By precisely controlling the growth conditions, the continuity of the newly added voxels and the fracture area is ensured, which effectively compensates for the volume measurement error caused by image noise or incomplete segmentation and improves the accuracy of volume calculation.

[0029] Advantages of additional aspects of the present invention will be given in part in the following description, and in part will become obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0031] Figure 1 A schematic diagram of the structure of the alveolar bone fracture detection system provided in Example 1 of the present invention; Figure 2This is a flow chart of the alveolar bone fracture detection method implemented by the alveolar bone fracture detection system provided in Example 1 of the present invention. DETAILED DESCRIPTION

[0032] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0033] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0034] It should be noted that the terms used herein are only for describing specific embodiments, and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that the terms "include" and "comprise" and any variation are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0035] In the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

[0036] Example 1 This embodiment provides a system for detecting alveolar bone fractures. Figure 1-Figure 2 As shown, including: A feature extraction module is configured to extract multi-scale features from the acquired oral image data, thereby segmenting the fracture area and the tooth root area; A type recognition module is configured to extract a center line of the fracture area, accumulate three-dimensional Euclidean distances voxel by voxel along the center line to obtain the length of the fracture line, express the center line as a function of the arc length and then calculate the global average curvature, detect the number of branch points in the fracture area, and calculate the minimum distance between the fracture area and the root area, thereby identifying the fracture type according to the length of the fracture line, the global average curvature, the number of branch points, and the minimum distance between the fracture area and the root area; The volume calculation module is configured to use the centroid of the fracture area as the seed point and determine the grayscale value of the seed point, merge the newly added voxels that belong to the neighborhood extension area of ​​the original fracture area and whose grayscale value difference with the grayscale value of the seed point meets the set threshold into the original fracture area, obtain the optimized fracture area, and calculate the regional volume of the optimized fracture area.

[0037] In this embodiment, the system also includes a data acquisition and preprocessing module, which acquires the patient's oral image data through a CBCT (Cone beam CT) device and performs preprocessing using an adaptive denoising algorithm to remove noise while retaining details, thereby ensuring the high quality of the oral image data and the accuracy of subsequent analysis.

[0038] The denoising optimization formula is: ; in, is the preprocessed oral image data; The original oral image data is in coordinates The pixel value at ; The preprocessed oral image data is in coordinates The pixel value at ; is a region defined on the oral image data, indicating the summation within the region; It is a balance parameter used to adjust the weight between denoising and detail preservation; The preprocessed oral image data is in coordinates The gradient at , indicates the degree of image change.

[0039] This embodiment introduces adaptive denoising technology in the denoising process, which can remove noise while retaining details, significantly improving image quality. Compared with traditional denoising methods, the method of this embodiment can better handle complex background noise, ensure that the image is more accurate in detail and contour, and provide a reliable image basis for subsequent feature extraction.

[0040] In this embodiment, in the feature extraction module, a segmentation model (Segment Anything Model, SAM) is used as a feature extraction network to perform fine segmentation and feature capture on the fracture area.

[0041] The SAM model consists of an encoder and a decoder. The encoder is responsible for extracting the multi-scale features of the input oral image data, and the decoder segmentates the fracture area based on the extracted multi-scale features.

[0042] The basic process is expressed as: ; in, It is The oral imaging data of the first step, and They are The weight parameters of the encoder and decoder, It is Step segmentation result.

[0043] The goal of the encoder is to extract multi-scale features of the input oral image data and capture hierarchical information from local details (such as the edge of the fracture line) to global semantics (such as the anatomical structure of the jaw). ResNet-50 is used as the encoder backbone network, and the hierarchical structure is shown in Table 1, where Conv is the convolution layer, MaxPool is the maximum pooling layer, and ResBlock is the residual block.

[0044] Table 1 Hierarchical structure; .

[0045] The residual block (ResBlock) structure is: ;in, It is a convolution operation, which consists of two 3×3 convolutional layers, each followed by a batch normalization layer and a ReLU activation function; is the input feature map, is the output feature map; Represents the set of weight parameters in the convolution operation, including the kernel weights and bias terms of each 3×3 convolution layer, which are used to input feature maps. Perform a convolution operation on it to generate .

[0046] The encoder outputs 5 scale feature maps (corresponding to stages 1-5), which are marked as: ;in, It is defined as a low-level feature, which retains detailed information such as fracture line edge and texture; It is defined as a high-level feature that encodes the overall morphology of the jaw and the semantic context of the fracture area.

[0047] The goal of the decoder is to gradually restore the spatial resolution, fuse the multi-scale features of the encoder, and generate a fracture area segmentation mask. Specifically, it includes: This embodiment introduces adaptive weight calculation in the feature extraction process of the SAM model to automatically adjust the importance of features in different regions. For each feature layer, the adaptive weight is: ; in, It is The first step of oral imaging data Layer feature map; and They are The mean and standard deviation of the layer feature map; is the smoothing factor; It is Step Adaptive weights of layer features, Reflection feature layer The significance of , a high weight means that the layer feature contributes more to the fracture line segmentation; for The spatial location of each voxel (or pixel) in ; Represents the feature map All spatial locations The voxels are traversed and summed.

[0048] Based on adaptive weights, multi-scale feature fusion is performed to combine features of different scales for information enhancement. Multi-scale feature fusion is expressed as: ; in, is the fusion feature, is the total number of feature layers.

[0049] Fusion Features Directly input into the decoder as the initial input feature of the decoder, and then use the Skip Connection method to connect the encoder stage Feature map Feature map of the layer corresponding to the decoder Stitching, gradually restore the spatial resolution and generate segmentation masks. Decoded as stage 4: ;in, Splicing for channel dimension; It is a transposed convolution with stride = 2 and kernel size = 3×3.

[0050] After the decoder generates the segmentation result, it fuses the semantic information and edge details of the features and then further optimizes them through the conditional random field. The conditional random field smoothes the segmentation edges by modeling the spatial relationship between pixels to ensure the continuity and accuracy of the fracture line.

[0051] In this embodiment, the advantages of this module are: Adaptive weight calculation mechanism: Introducing an adaptive feature extraction mechanism, when processing each layer of features, the weights are automatically adjusted according to the significance of different regions. This mechanism can optimize the model according to the characteristics of different regions in the oral imaging data, making the detection of key areas (such as fracture lines) more sensitive, avoiding the excessive focus on unimportant areas in traditional methods.

[0052] Multi-scale feature fusion: By introducing a multi-scale feature fusion strategy, the low-level edge information and high-level semantic information are effectively combined, which not only improves the accuracy and robustness of image segmentation, but also enhances the model's ability to handle complex images (such as noise interference or blurred edge areas), thereby improving the diagnostic accuracy of fracture lines.

[0053] In this embodiment, after the fracture area is segmented, the conditional random field (CRF) is used to optimize the segmentation result to improve the accuracy of the segmentation edge and the reliability of the overall diagnosis. The basic principle is to regard the segmentation result as a node and optimize the segmentation by establishing the relationship between the nodes.

[0054] The CRF energy function is: ; in, is the output label of the model; is the input image; is the cost of a single node (i.e., the segmentation confidence for each pixel); is the smoothing cost between adjacent nodes, which can ensure the classification consistency of adjacent pixels.

[0055] Single point cost calculation: The single point cost of a node is calculated based on the pixel value and classification probability of the segmented image, specifically: ;in, Pixel The probability of being classified into a certain category.

[0056] Double-point cost calculation: Double-point cost is used to enhance the smoothness of the segmentation boundary, usually implemented by Gaussian kernel function: ;in, Pixel and The feature vector of . is a parameter that controls the degree of smoothing.

[0057] Parameter optimization: The iterative conditional mode (ICM) algorithm is used to optimize the energy function and iteratively update the category label of each pixel until convergence. In each iteration, the optimization goal is to reduce the energy function value: ;in, It is Step The label of the iteration, It is Step The label of the iteration, is the learning rate, It is The energy function gradient of the iteration. Through continuous iteration, the label prediction is optimized to ensure that the segmentation result is accurate and consistent with the feature extraction in the previous stage.

[0058] In this embodiment, the conditional random field is used to post-process the preliminary segmentation results to further optimize the segmentation edges. The CRF model can smooth adjacent pixels and enhance the continuity of edges by modeling the spatial relationship between pixels, making the segmentation results more refined and realistic.

[0059] In this embodiment, the type recognition module realizes automatic classification of alveolar bone fracture types based on the morphological characteristics, spatial distribution and relationship of the fracture line with surrounding anatomical structures (such as tooth roots).

[0060] Specifically include: (1) Extraction of fracture line morphological features.

[0061] (1-1) Fracture line skeletonization and centerline extraction.

[0062] Binary mask of fracture area after CRF optimization (1 represents the fracture area, 0 represents the normal area), extract the centerline coordinates of the fracture area : .

[0063] Skeletonization is an image processing technique used to simplify the fracture area to its centerline while preserving its basic shape and topology. Skeletonization and skeleton point identification are based on the fracture area binary mask after CRF optimization. achieved; Specifically, it includes: calculating the distance from the point in the fracture area to the nearest boundary, and based on the distance result, extracting the skeleton through iterative corrosion operation, which gradually reduces the area until only the center line remains. ,Each skeleton point is a coordinate point on the center line of the fracture line, representing a key position of the fracture line.

[0064] (1-2) Calculation of characteristic parameters.

[0065] (a) Calculate the fracture line length by accumulating the 3D Euclidean distance voxel by voxel along the center line Its physical meaning is to reflect the spatial extension of the fracture line. If it is larger than the set threshold, it is considered a long fracture line; ; Among them, voxel is the abbreviation of volume pixel. For the The coordinates of the centerline voxels; For the The coordinates of the centerline voxels; is the total prime number of the center line.

[0066] (b) Calculate the global mean curvature , to reflect the local curvature of the center line, the average curvature If it is greater than the set threshold (for example, C>0.5), it is considered a high curvature value, indicating that there is a sharp turn in the fracture line, which is common in comminuted fractures.

[0067] Specifically include: Expressing the center line as arc length Function to parameterize the center line: ; The curvature formula is: ;in, is the first-order derivative (tangent direction); is the second-order derivative (direction of curvature); The global mean curvature is: .

[0068] (c) Traverse the skeleton points and determine whether there are connected directions greater than the set threshold in the neighborhood of the current skeleton point. If so, the current skeleton point is regarded as a branch point to detect the number of branch points in the fracture area. , This indicates that the fracture line has multiple bifurcations, which is a typical feature of a comminuted fracture.

[0069] Specifically, for each skeleton point, check other points within the set range around the skeleton point (i.e., the neighborhood). If there are 3 or more directions extending from this skeleton point in the neighborhood (i.e., there are also fracture lines in these directions), then the skeleton point is considered to be a branch point.

[0070] (d) By traversing all fracture area voxels and root region voxels , calculate the Euclidean distance between the two, and finally take the minimum value as the minimum distance between the fracture area and the root area ; That is, by calculating the fracture area and the root area The minimum distance ,like Less than the set distance threshold (such as ), it indicates that the fracture involves the tooth root and needs to be additionally marked as "periradicular fracture"; ; in, The fracture area The three-dimensional coordinates of a voxel represent the actual spatial position of the voxel in the image data; The root area The three-dimensional coordinates of a voxel represent the spatial position of a specific voxel in the tooth root area.

[0071] (2) Classification logic and threshold setting.

[0072] (a) If the fracture line length is greater than or equal to the set length threshold, the global average curvature is less than or equal to the set first curvature threshold, and the number of branch points is less than or equal to the set first number threshold (for example ), it is a linear fracture; its clinical basis is that it is long in length, low in curvature and has no bifurcation, which conforms to the morphological characteristics of a linear fracture.

[0073] (b) If the number of branch points is greater than or equal to the set second number threshold, or the global average curvature is greater than the set second curvature threshold (for example ), it is a comminuted fracture; the clinical basis is that multiple branches or high curvature indicate that the fracture line is complex and meets the definition of comminuted fracture.

[0074] (c) If the minimum distance between the fracture area and the root area is less than the distance threshold (e.g. ), it is a periradicular fracture; its clinical basis is that the fracture involves the tooth root, which meets the definition of periradicular fracture.

[0075] In this embodiment, when calculating the fracture area volume When performing volumetric analysis, the region growing algorithm is first used to further refine the voxels to ensure the accuracy of volume measurement.

[0076] Specific: Seed selection: using the centroid coordinates of the original fracture area is the seed point to ensure that the algorithm focuses on the core area; the centroid coordinates are: ;in, are the coordinates of all fracture voxels in the fracture region binary mask, is the total number of all fracture voxels in the binary mask of the fracture area.

[0077] Added Voxels The conditions for being merged must be met at the same time: ;in, is the gray value of the seed point (consistent with the typical value of the fracture area, such as 200-400 HU); , represents the tolerance. In CBCT images, the grayscale value of the fracture area is usually within a certain range (such as 200-400 HU), and the tolerance range of 50 HU can cover the grayscale fluctuation caused by image noise or partial volume effect. It is set based on clinical experience and experimental verification, which can balance sensitivity and specificity, and will neither miss the real fracture voxels nor mistakenly include non-fracture voxels; neighborhood Represents the neighborhood expansion area of ​​the original fracture area binary mask to ensure that the newly added voxels are continuous with the fracture area.

[0078] Therefore, combined with the regional growth, the volume of the new voxels is increased , the fracture area volume is calculated for: ; ; in, The first The cross-sectional area of ​​the layer voxel is obtained by converting the pixel count to the physical size of a single pixel; is the CBCT inter-slice distance (derived from data acquisition); is the volume of the newly added voxels for regional growth, M is the number of newly added voxels, It is the volume error compensation part; is the width of the new voxel, is the height of the newly added voxel, To add the voxel layer spacing, for isotropic voxels, , that is, each voxel has the same size in three directions.

[0079] In this embodiment, the actual spatial volume is directly calculated based on the three-dimensional voxel model of the binary mask of the fracture area, avoiding the error caused by projection superposition in traditional two-dimensional measurement. It supports volume integration of complex fracture morphology (such as comminuted fractures in multiple areas), and the volume error compensation part can dynamically compensate for the missed volume of scattered fracture fragments.

[0080] In this embodiment, a region growing algorithm is used for voxel refinement. By precisely controlling the growth conditions, the continuity of the newly added voxels and the fracture area is ensured, the volume measurement errors caused by image noise or incomplete segmentation are effectively compensated, and the accuracy of volume calculation is improved.

[0081] In this embodiment, by extracting the centroid coordinates and the volume of the lesion area from the segmentation results, the doctor can be provided with accurate fracture location and quantified volume data, which can not only help the doctor quickly understand the severity of the fracture, but also provide quantitative support for treatment decisions. And a complete diagnostic report is generated based on the automatic diagnosis results, including detailed information such as fracture type, location and volume. By automatically generating reports, the doctor's workload is reduced and the accuracy and consistency of the reports are ensured.

[0082] The alveolar bone fracture detection system based on intelligent algorithms proposed in this embodiment realizes efficient and accurate diagnosis and quantitative analysis of alveolar bone fractures through deep fusion of deep learning, three-dimensional reconstruction and adaptive optimization technology. With CBCT images as input, the improved SAM model combined with the conditional random field optimization algorithm is used to complete the refined segmentation of fracture lines; through the joint analysis of multi-scale morphological features (length, curvature, number of branches) and spatial relationships (root distance), the automatic classification of linear fractures, comminuted fractures and periradicular fractures is realized; and based on the three-dimensional mask and dynamic region growing algorithm, the high-precision measurement of fracture volume is completed.

[0083] Compared with the traditional diagnosis method that relies on the doctor's experience, the system in this embodiment has the following core advantages: Full process automation: from image preprocessing to report generation, reducing manual intervention and improving diagnostic efficiency.

[0084] Multi-dimensional quantitative output: Provides fracture type, location coordinates, volume percentage and morphological parameters, providing quantitative support for surgical planning and prognosis assessment.

[0085] Clinical interpretability: All diagnostic results are based on traceable morphological and imaging features (such as curvature calculation and grayscale constraints), which are consistent with the doctor's cognitive logic.

[0086] Enhanced robustness: Adaptive denoising, region growing compensation and other mechanisms can significantly suppress the influence of image noise and partial volume effect.

[0087] It should be noted that all data is obtained in compliance with laws and regulations and user consent, and the data is used legally.

[0088] In further embodiments, there is also provided: An electronic device includes a memory and a processor, and computer instructions stored in the memory and running on the processor, wherein when the computer instructions are run by the processor, a method for detecting alveolar bone fracture is completed, and the method for detecting alveolar bone fracture includes: Extract multi-scale features from the acquired oral image data, thereby segmenting the fracture area and the tooth root area; Extract the center line of the fracture area, accumulate the three-dimensional Euclidean distance voxel by voxel along the center line to obtain the length of the fracture line, express the center line as a function of the arc length, calculate the global average curvature, detect the number of branch points in the fracture area, and calculate the minimum distance between the fracture area and the root area. In this way, the fracture type can be identified based on the length of the fracture line, the global average curvature, the number of branch points, and the minimum distance between the fracture area and the root area. The centroid of the fracture area is taken as the seed point, and the grayscale value of the seed point is determined. The newly added voxels that belong to the neighborhood extension area of ​​the original fracture area and whose grayscale value difference with the grayscale value of the seed point meets the set threshold are merged into the original fracture area to obtain the optimized fracture area, and the regional volume of the optimized fracture area is calculated.

[0089] It should be understood that in this embodiment, the processor may be a central processing unit CPU, and the processor may also be other general-purpose processors, digital signal processors DSP, application-specific integrated circuits ASIC, off-the-shelf programmable gate arrays FPGA or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0090] The memory may include a read-only memory and a random access memory, and provide instructions and data to the processor. A portion of the memory may also include a non-volatile random access memory. For example, the memory may also store information about the device type.

[0091] A computer-readable storage medium is used to store computer instructions. When the computer instructions are executed by a processor, a method for detecting alveolar bone fracture is completed. The method for detecting alveolar bone fracture includes: Extract multi-scale features from the acquired oral image data, thereby segmenting the fracture area and the tooth root area; Extract the center line of the fracture area, accumulate the three-dimensional Euclidean distance voxel by voxel along the center line to obtain the length of the fracture line, express the center line as a function of the arc length, calculate the global average curvature, detect the number of branch points in the fracture area, and calculate the minimum distance between the fracture area and the root area. In this way, the fracture type can be identified based on the length of the fracture line, the global average curvature, the number of branch points, and the minimum distance between the fracture area and the root area. The centroid of the fracture area is taken as the seed point, and the grayscale value of the seed point is determined. The newly added voxels that belong to the neighborhood extension area of ​​the original fracture area and whose grayscale value difference with the grayscale value of the seed point meets the set threshold are merged into the original fracture area to obtain the optimized fracture area, and the regional volume of the optimized fracture area is calculated.

[0092] The method in Example 1 can be directly embodied as a hardware processor, or a combination of hardware and software modules in the processor. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware. To avoid repetition, it is not described in detail here.

[0093] A computer program product, comprising a computer program, wherein when the computer program is executed by a processor, a method for detecting alveolar bone fracture is implemented, wherein the method for detecting alveolar bone fracture comprises: Extract multi-scale features from the acquired oral image data, thereby segmenting the fracture area and the tooth root area; Extract the center line of the fracture area, accumulate the three-dimensional Euclidean distance voxel by voxel along the center line to obtain the length of the fracture line, express the center line as a function of the arc length, calculate the global average curvature, detect the number of branch points in the fracture area, and calculate the minimum distance between the fracture area and the root area. In this way, the fracture type can be identified based on the length of the fracture line, the global average curvature, the number of branch points, and the minimum distance between the fracture area and the root area. The centroid of the fracture area is taken as the seed point, and the grayscale value of the seed point is determined. The newly added voxels that belong to the neighborhood extension area of ​​the original fracture area and whose grayscale value difference with the grayscale value of the seed point meets the set threshold are merged into the original fracture area to obtain the optimized fracture area, and the regional volume of the optimized fracture area is calculated.

[0094] The present invention also provides at least one computer program product tangibly stored on a non-transitory computer-readable storage medium. The computer program product includes computer executable instructions, such as instructions included in a program module, which are executed in a device on a real or virtual processor of the target to perform the process / method as described above. Typically, program modules include routines, programs, libraries, objects, classes, components, data structures, etc. that perform specific tasks or implement specific abstract data types. In various embodiments, the functions of program modules can be combined or divided between program modules as needed. Machine executable instructions for program modules can be executed in local or distributed devices. In distributed devices, program modules can be located in local and remote storage media.

[0095] The computer program code for implementing the method of the present invention can be written in one or more programming languages. These computer program codes can be provided to the processor of a general-purpose computer, a special-purpose computer or other programmable data processing device, so that the program code, when executed by the computer or other programmable data processing device, causes the function / operation specified in the flow chart and / or block diagram to be implemented. The program code can be executed completely on a computer, partially on a computer, as an independent software package, partially on a computer and partially on a remote computer or completely on a remote computer or server.

[0096] In the context of the present invention, computer program codes or related data may be carried by any appropriate carrier to enable a device, apparatus or processor to perform the various processes and operations described above. Examples of carriers include signals, computer readable media, etc. Examples of signals may include electrical, optical, radio, acoustic or other forms of propagation signals, such as carrier waves, infrared signals, etc.

[0097] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with this embodiment can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0098] Although the above describes the specific implementation mode of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without creative work are still within the scope of protection of the present invention.

Claims

1. A system for detecting alveolar bone fracture, characterized in that: include: A feature extraction module is configured to extract multi-scale features from the acquired oral image data, thereby segmenting the fracture area and the tooth root area; A type recognition module is configured to extract a center line of the fracture area, accumulate three-dimensional Euclidean distances voxel by voxel along the center line to obtain the length of the fracture line, express the center line as a function of the arc length and then calculate the global average curvature, detect the number of branch points in the fracture area, and calculate the minimum distance between the fracture area and the root area, thereby identifying the fracture type according to the length of the fracture line, the global average curvature, the number of branch points, and the minimum distance between the fracture area and the root area; The volume calculation module is configured to use the centroid of the fracture area as the seed point and determine the grayscale value of the seed point, merge the newly added voxels that belong to the neighborhood extension area of ​​the original fracture area and whose grayscale value difference with the grayscale value of the seed point meets the set threshold into the original fracture area, obtain the optimized fracture area, and calculate the regional volume of the optimized fracture area.

2. The alveolar bone fracture detection system according to claim 1, characterized in that: In the feature extraction module, during the process of extracting multi-scale features, adaptive weights are calculated for each feature layer: ； in, It is The first step of oral imaging data Layer feature map; and They are The mean and standard deviation of the layer feature map; is the smoothing factor; It is Step Adaptive weights of layer features; for The spatial position of each voxel in ; Multi-scale feature fusion based on adaptive weights: ； in, is the fusion feature, is the total number of feature layers.

3. The alveolar bone fracture detection system according to claim 1, characterized in that: In the type recognition module, the fracture line length for: ； in, For the The coordinates of the centerline voxels; For the The coordinates of the centerline voxels; is the total prime number of the center line.

4. The alveolar bone fracture detection system according to claim 1, characterized in that: In the type recognition module, the process of calculating the global mean curvature includes: representing the center line as an arc length Function , the global mean curvature for: ； in, is the first-order derivative of the function; is the second-order derivative of the function; is the total prime number of the center line.

5. The alveolar bone fracture detection system according to claim 1, characterized in that: In the type recognition module, the process of detecting the number of branch points in the fracture area includes: traversing the skeleton points on the center line, determining whether there are connection directions greater than a set threshold number in the neighborhood of the current skeleton point, and if so, treating the current skeleton point as a branch point, thereby determining the number of branch points in the fracture area.

6. The alveolar bone fracture detection system according to claim 1, characterized in that: In the type identification module, the process of identifying the fracture type includes: If the fracture line length is greater than or equal to the set length threshold, the global average curvature is less than or equal to the set first curvature threshold, and the number of branch points is less than or equal to the set first number threshold, then it is a linear fracture; If the number of branch points is greater than or equal to the set second number threshold, or the global average curvature is greater than the set second curvature threshold, it is a comminuted fracture; If the minimum distance between the fracture area and the root area is less than the distance threshold, it is a periradicular fracture.

7. The alveolar bone fracture detection system according to claim 1, characterized in that: In the volume calculation module, the centroid coordinates are ;in, are the coordinates of all fracture voxels in the fracture region binary mask, is the total number of all fracture voxels in the fracture area; Added Voxels The conditions for being merged are: ;in, is the gray value of the seed point, Tolerance, Neighborhood represents the neighborhood extension area of ​​the original fracture area; Area Volume for: ； in, The first The cross-sectional area of ​​the layer voxel; is the distance between oral image data layers; is the volume of the newly added voxels, and M is the number of newly added voxels.

8. An electronic device, characterized in that: The invention comprises a memory and a processor, and computer instructions stored in the memory and executed on the processor. When the computer instructions are executed by the processor, a method for detecting alveolar bone fracture is completed. The method for detecting alveolar bone fracture comprises: Extract multi-scale features from the acquired oral image data, thereby segmenting the fracture area and the tooth root area; Extract the center line of the fracture area, accumulate the three-dimensional Euclidean distance voxel by voxel along the center line to obtain the length of the fracture line, express the center line as a function of the arc length, calculate the global average curvature, detect the number of branch points in the fracture area, and calculate the minimum distance between the fracture area and the root area. In this way, the fracture type can be identified based on the length of the fracture line, the global average curvature, the number of branch points, and the minimum distance between the fracture area and the root area. The centroid of the fracture area is taken as the seed point, and the grayscale value of the seed point is determined. The newly added voxels that belong to the neighborhood extension area of ​​the original fracture area and whose grayscale value difference with the grayscale value of the seed point meets the set threshold are merged into the original fracture area to obtain the optimized fracture area, and the regional volume of the optimized fracture area is calculated.

9. A computer-readable storage medium, characterized in that: Used to store computer instructions, when the computer instructions are executed by the processor, the alveolar bone fracture detection method is completed, and the alveolar bone fracture detection method includes: Extract multi-scale features from the acquired oral image data, thereby segmenting the fracture area and the tooth root area; Extract the center line of the fracture area, accumulate the three-dimensional Euclidean distance voxel by voxel along the center line to obtain the length of the fracture line, express the center line as a function of the arc length, calculate the global average curvature, detect the number of branch points in the fracture area, and calculate the minimum distance between the fracture area and the root area. In this way, the fracture type can be identified based on the length of the fracture line, the global average curvature, the number of branch points, and the minimum distance between the fracture area and the root area. The centroid of the fracture area is taken as the seed point, and the grayscale value of the seed point is determined. The newly added voxels that belong to the neighborhood extension area of ​​the original fracture area and whose grayscale value difference with the grayscale value of the seed point meets the set threshold are merged into the original fracture area to obtain the optimized fracture area, and the regional volume of the optimized fracture area is calculated.

10. A computer program product, characterized in that The invention comprises a computer program, wherein when the computer program is executed by a processor, a method for detecting alveolar bone fracture is implemented, wherein the method for detecting alveolar bone fracture comprises: Extract multi-scale features from the acquired oral image data, thereby segmenting the fracture area and the tooth root area; Extract the center line of the fracture area, accumulate the three-dimensional Euclidean distance voxel by voxel along the center line to obtain the length of the fracture line, express the center line as a function of the arc length, calculate the global average curvature, detect the number of branch points in the fracture area, and calculate the minimum distance between the fracture area and the root area. In this way, the fracture type can be identified based on the length of the fracture line, the global average curvature, the number of branch points, and the minimum distance between the fracture area and the root area. The centroid of the fracture area is taken as the seed point, and the grayscale value of the seed point is determined. The newly added voxels that belong to the neighborhood extension area of ​​the original fracture area and whose grayscale value difference with the grayscale value of the seed point meets the set threshold are merged into the original fracture area to obtain the optimized fracture area, and the regional volume of the optimized fracture area is calculated.

Images